Ebola Virus Exploits a Monocyte Differentiation Program To Promote Its Entry

Abstract

Antigen-presenting cells (APCs) are critical targets of Ebola virus (EBOV) infection in vivo. However, the susceptibility of monocytes to infection is controversial. Studies indicate productive monocyte infection, and yet monocytes are also reported to be resistant to EBOV GP-mediated entry. In contrast, monocyte-derived macrophages and dendritic cells are permissive for both EBOV entry and replication. Here, freshly isolated monocytes are demonstrated to indeed be refractory to EBOV entry. However, EBOV binds monocytes, and delayed entry occurs during monocyte differentiation. Cultured monocytes spontaneously downregulate the expression of viral entry restriction factors such as interferon-inducible transmembrane proteins, while upregulating the expression of critical EBOV entry factors cathepsin B and NPC1. Moreover, these processes are accelerated by EBOV infection. Finally, ectopic expression of NPC1 is sufficient to rescue entry into an undifferentiated, normally nonpermissive monocytic cell line. These results define the molecular basis for infection of APCs and suggest means to limit APC infection.

INTRODUCTION

Zaire Ebola virus (EBOV) is an emerging zoonotic pathogen that has caused outbreaks of viral hemorrhagic fever in humans with fatality rates approaching 90% (1). EBOV tropism toward antigen-presenting cells (APCs) is thought to play an important role in viral pathogenesis, contributing to the establishment of infection and to the development of hemorrhagic fever (2). EBOV productively infects APCs, including monocytes, macrophages, and dendritic cells (DCs), in vitro (3–9), and tissue sections from EBOV-infected humans and nonhuman primates contain APCs positive for EBOV antigen/nucleic acid, demonstrating APC infection in vivo (10–16). Although APCs serve as sites of virus amplification, their infection also deregulates APC function (2, 17, 18). This deregulation may contribute to the development of an ineffective antiviral immune response, as well as an intense and deregulated inflammatory response (4, 19).

Although monocytes, the most abundant blood-borne APCs, likely contribute to EBOV pathogenesis, an apparent discrepancy exists in our understanding of monocyte infection by EBOV (18). Specifically, although EBOV productively infects human blood-derived monocytes (4, 20), several other studies demonstrate limited or restricted EBOV GP-mediated entry into monocytes (21–23).

EBOV entry, which includes attachment and penetration into the target cell cytoplasm, is mediated by the surface glycoprotein (GP) (24). GP likely mediates viral attachment through the receptor-binding domain (RBD) located at its N terminus (25–28). Subsequent viral uptake likely occurs via macropinocytosis (29–32). A number of cell surface molecules, including C-type lectins, have been identified as attachment or entry factors (33–38). However, none of these factors appears to function as an essential cell surface receptor. Additional host factors have also been implicated as regulating entry, including components of the homotypic fusion and vacuole protein-sorting (HOPS) multisubunit tethering complex, the cysteine proteases cathepsin L and cathepsin B and signaling molecules, including acid sphingomyelinase and phosphoinositide-3 kinase (25, 26, 39–44). Recently, Niemann-Pick C1 (NPC1), an endosomal protein involved in cholesterol transport and storage, was identified as an essential EBOV entry receptor (39, 45, 46). Negatively acting entry restriction factors have also been identified, including the interferon-inducible transmembrane proteins (IFITMs) (47, 48).

Here we sought to define the basis of EBOV tropism toward monocytes, macrophages, and DCs. The data indicate that undifferentiated monocytes are indeed refractory to EBOV entry, whereas macrophages and DCs are fully permissive. However, EBOV can associate with undifferentiated monocytes and can complete the entry process as the cells spontaneously differentiate into macrophages or DCs. This results in substantially delayed entry kinetics and less robust cytokine responses of monocytes relative to differentiated macrophages or DCs. Profiling an array of genes previously implicated in EBOV entry demonstrates that during monocyte differentiation essential entry factors NPC1 and cathepsin B increase, while the restriction factors IFITM1, -2, and -3 decrease. Furthermore, although monocyte infection with EBOV accelerates the rate of IFITM downregulation, the kinetics of upregulation of cathepsin B and NPC1 expression remain largely unchanged. Lastly, overexpression of NPC1 in THP-1 monocytes, which are nonpermissive for EBOV GP-mediated entry unless differentiated into macrophage-like cells, partially rescued EBOV GP-mediated entry.

MATERIALS AND METHODS

Cell culture, vectors, and plasmids.

HEK293T (293T) cells were cultured in Dulbecco modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% l-glutamine. The plasmids used in transfections included (i) pcDNA BLA-VP40, which expresses a β-lactamase-Zaire EBOV VP40 chimera (49, 50), (ii) pCAGGS GFP-VP40, which expresses a VP40-GFP chimera (51), (iii) pcDNA EBGP, which expresses wild-type Zaire EBOV strain Mayinga GP, (iv) pcDNA EBGPF88A (50, 52, 53), (v) pcDNA EBGPF159A (53), or (vi) pcDNA VSVG, which expresses vesicular stomatitis virus glycoprotein (VSV G). NPC1 was expressed using the pBABE expression retroviral vector kindly provided by Kartik Chandran (Albert Einstein College of Medicine).

Isolation and culture of human monocytes, macrophages, and DCs.

Peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll density gradient centrifugation (Histopaque; Sigma-Aldrich, St. Louis, MO) from the buffy coats of healthy human donors (New York Blood Center). CD14⁺ monocytes were purified using anti-human CD14 antibody-labeled magnetic beads and iron-based LS columns (Miltenyi Biotec, Auburn, CA) and used immediately or further differentiated into macrophages or DCs. Stains of isolated monocytes typically showed >90% positivity for CD14 staining. Monocytes were plated (0.7 × 10⁶ to 1 × 10⁶ cells/ml) in DC medium (RPMI [Invitrogen, Carlsbad, CA] supplemented with 100 U of penicillin/ml, 10 μg of streptomycin/ml, 55 μM β-mercaptoethanol, and 4% human serum AB [GemCell; Gemini Bio-Products, West Sacramento, CA]) supplemented with either 500 U of human granulocyte-macrophage colony-stimulating factor (GM-CSF; Peprotech, Rocky Hill, NJ)/ml, 500 U of human interleukin-4 (IL-4; Peprotech)/ml to differentiate into DCs (54), or 20 to 50 ng of macrophage colony-stimulating factor (M-CSF; Peprotech)/ml to differentiate into macrophages. By day 5, immature DCs expressed surface CD11c, CD1c, and HLA-DR (major histocompatibility complex [MHC] II) and low/negative levels of CD14, whereas macrophages retained high levels of CD14 and were HLA-DR positive (55, 56) and were permissive for EBOV entry and infection (50).

VLP production and purification.

For virus-like particle (VLP) production, 293T cells were cotransfected with a combination of two plasmids using Lipofectamine 2000 (DNA/Lipofectamine 2000 ratio of 1:1): either pcDNA BLA-VP40 or pCAGGS GFP-VP40, and either pcDNA expressing EBGP or VSV G at a ratio of 3:2 for VP40 to envelope glycoprotein. The VLPs were harvested at 3 days posttransfection; the 293T cell spent supernatant was layered over 10 ml of 20% sucrose in NTE buffer (100 mM NaCl, 20 mM Tris-hydrochloride [pH 7.5], and 1 mM EDTA [EDTA]), and VLPs were pelleted at 25,000 rpm in an SW-28 rotor (∼80,000 × g) for 2 h at 4°C. The VLPs were then gently washed without resuspension with 25 ml of cold NTE or phosphate-buffered saline (PBS) and centrifuged again. The VLPs were finally resuspended in a total of 150 μl of NTE and stored on ice at 4°C.

Flow cytometry.

Flow cytometry was performed using an LSR II flow cytometer equipped with a violet laser (BD Bioscience). The data were analyzed using FlowJo software (Tree Star, Ashland, OR).

Mouse macrophages and dendritic cells.

C57/BL6 mice were sacrificed in order to obtain peritoneal macrophages (large CD11b⁺ cells) or bone marrow cells according to standard techniques. In order to induce monocyte differentiation into dendritic cells (DCs) or macrophages, mouse bone-marrow cells grown in DMEM plus 10% FBS were supplemented with mouse GM-CSF (50 ng/ml; Peprotech) or M-CSF (20 ng/ml; Peprotech), respectively. All animal procedures were performed in accordance with Institutional Animal Care and Use Committee (IACUC) guidelines and have been approved by the IACUC of Mount Sinai School of Medicine.

Determining EBOV infectious titers.

EBOV titers were determined using an agarose plaque assay. Briefly, VeroE6 cells were plated in six-well plates to 90 to 100% confluence. Fivefold dilutions of stock virus were prepared (10⁻¹ to 10^−6.5). Medium samples were removed from the six-well plates, and 200 μl of inoculum for each dilution was added to each well of the plate (six replicates), rocked, and allowed to incubate for 1 h at 37°C and 5% CO₂, with rocking every 15 min to prevent the cell layer from drying. At the end of the absorption period, each well was overlaid with 2 ml of a 1× agarose overlay (1% agarose mixed with 2× Eagle modified essential medium [EMEM; Gibco], 2× Anti-Anti [Gibco], 2× GlutaMAX [Gibco], and 10% fetal calf serum [FCS; Gibco] to yield a final overlay of 0.5% agarose [SeaKem ME agarose; Lonza], 1× EMEM, 1× Anti-Anti, 1× GlutaMAX, and 5% FCS). The agarose was allowed to solidify, and the plates were returned to an incubator. On day 9 after the primary overlay, a secondary overlay of the primary overlay mixture plus 4% neutral red (Gibco) was added, the samples were allowed to solidify, and the cells were again returned to the incubator for an additional night. Plaques were counted the following day, and the titers were determined.

Entry assays.

Entry assays were performed as outlined in a previous study (50). Different VLP preparations were used such that equivalent levels of β-lactamase activity were present for the entry assays. VLPs were mixed with target cells and centrifuged (“spinoculated”) at 1,850 rpm for 45 min at 4°C before incubation at 37°C in 2% RPMI medium for the indicated times. Target cells were then loaded with fluorescent CCF2-AM substrate for 1 h at room temperature. An LSR II (BD Bioscience) flow cytometer equipped with a violet laser was used to excite the CCF2-AM substrate, and events (cells) were assayed for green and blue fluorescence after sorting for live cells. Live cells were sorted using side-scatter and forward-scatter properties (low side scatter and high forward scatter [data not shown]) and for their ability to retain the CCF2-AM substrate (cells that fluoresce green) since at least a subset of dead cells are permeable and therefore leaky. Cells fluorescing blue compared to the control (for example, mock control [results not shown]) were scored as entry positive.

EBOV infections.

All EBOV infections were performed at the U.S. Army Medical Research Institute of Infectious Diseases at biosafety level 4. Monocytes, macrophages, and DCs were infected with EBOV_GFP (50) or EBOV (Kikwit 1995) at the indicated multiplicities of infection (MOIs) for 1 to 2 h with rocking. When infecting human cells with the EBOV_GFP, we used MOIs of 5 and 10 because this consistently results in a significant level of infection (>20%). After infection, cells were washed three times with PBS, resuspended, and plated in triplicate. To enumerate the percentages of cells infected by EBOV_GFP, the cells were gently scraped off the plates on at 2 and 3 days postinfection and analyzed by flow cytometry after gating out dead cells using side-scatter and forward-scatter properties (side scatter low, forward scatter high). To identify secreted cytokines induced after EBOV_GFP infection, day 3 spent supernatants from infected and noninfected control cells were analyzed as described below. To determine the gene expression (mRNA) levels, total RNA was first harvested using Tri-Reagent, which was added to each well containing (infected) cells, and then pipetted to make sure the cells were dissociated and lysed, and the cell lysate was further processed as outlined below.

RNA isolation and cDNA synthesis.

RNA was extracted from the aqueous phase of the Tri-Reagent (Sigma-Aldrich) samples using an RNeasy micro kit (Qiagen), including a DNase step as recommended by the manufacturer. cDNA was generated from isolated mRNA using a Superscript III first-strand synthesis kit (Invitrogen) using an oligo(dT) primer as outlined by the manufacturer.

Quantitative reverse transcription-PCR (RT-PCR).

Primers were designed by using online real-time PCR primer design software and selecting intron-spanning primers (Roche Applied Science [http://www.rocheappliedscience.com/sis/rtpcr/upl/index.jsp?id=uplct_000000]). Since IFITM genes show high sequence identity, rather than just relying on the PCR primer design software, the IFITM primers were modified to span nucleotides that allowed for specific IFITM type amplification. EBOV NP primers were previously used to assay for EBOV NP expression (57). Table 1 provides the nucleotide sequences (5′ to 3′) for the forward and reverse primers used, respectively, for each human and EBOV gene.

Table 1.

Primersused for quantitative RT-PCR

Entry factor	Forward and reverse primer sequences (5′–3′)
IFITM1	CAACACCCTCTTCTTGAACTGG and GCCGAATACCAGTAACAGGATG
IFITM2	GTTCAACACCCTCTTCATGAACA and GACGACCAACACTGGGATGAT
IFITM3	CAACACCCTCTTCATGAACC and CCTAGACTTCACGGAGTAGG
VPS33A	GATCCGAGATAAGAACTTCAACG and GATCTCCCCCACGGTCTTAG
VPS11	TCGAATCTTCCCTGCTATTCC and AACACTGCCATCTGTGAAACC
VPS18	TTGTCGTCTCCAGCAATCAG and CCTTGCCCAAGTCAATGC
VPS41	GAATTTTTGGAGAGTGATTATGAGG and CGAACTGCTTGAACTATGACTGA
CTSB	CTGTGGCAGCATGTGTGG and CCTTTTCTTGTCCAGAAGTTCC
BLOC1S1	CGCCTCCTAAAAGAACACCA and TAGCCTCTCGCCTCCTCTTT
GNPTAB	TGAAGGAACTACAGCAGGTCAG and TGTGTTTTTCCCAAGGATTTCT
MGL	TACACCTGGATGGGCCTCAG and TGTTCCATCCACCCACTTCC
DC-SIGN	AGGCTCTGATTTCACGTCCC and TGGGGACCTTGGACACTG
NPC1	GGTCCGCCTGTGTACTTTGT and GGCTTCACCCAGTCGAAATA
LSECTIN	ATTCTGAGTATCCTATTGTCCAAGG and CTGCTTCGAGGCGTTTGT
DTK	CTACCTCATTGGCGGGAAC and ACTCCAGCACTGGTACATGAGA
MER	ATTGGAGACAGGACCAAAGC and GGGCAATATCCACCATGAAC
AXL	CGTAACCTCCACCTGGTCTC and TCCCATCGTCTGACAGCA
ITGB1	AATGAATGCCAAATGGGACACGGG and TTCAGTGTTGTGGGATTTGCACGG
ITGB2	AAACAACATCCAGCCCATCTTCGC and TGGACCACATTGCTGGAGT
TIM4	TGTGGAAAACGAGTGATTCTGT and TTGTTCTGCTCAGGAACTGC
GAPDH	CCATGTTCGTCATGGGTGTG and GGTGCTAAGCAGTTGGTGGTG
L SIGN	GCAGATACATGGCCACAAGA and GGACACTTGGACAAGGATGG
EBOV NP 2034f-2106r	CAGTGCGCCACTCACGGACA and TGGTGTCAGCATGCGAGGGC

Relative gene expression was determined using the iQ SYBR green Supermix (Bio-Rad) according to the manufacturer's instructions. The PCR temperature profile was 95°C for 10 min, followed by 40 cycles of 95°C for 10 s and 60°C for 60 s. All of the reactions were performed in triplicate, and standard error bars were added to graphs (see Fig. 7). CXF Manager software (Bio-Rad) was used to analyze the relative mRNA expression levels by the change in threshold cycle (ΔC_T) method using the GAPDH (glyceraldehyde-3-phosphate dehydrogenase) gene to normalize the results. To determine the NP mRNA levels in EBOV-infected monocytes and DCs, a standard curve generated by using an NP expression plasmid as a template was used to calculate the relative copy numbers of NP mRNA from equal numbers of cells.

Fig 7.

mRNA levels of factors that regulate EBOV entry in plated monocytes with or without EBOV infection. Freshly isolated blood monocytes were plated or infected with EBOV at an MOI of 1 before plating. RNA from infected and noninfected monocytes was harvested at 0, 3, 6, 12, 24, and 48 postinfection. Poly(A)-derived cDNA was used as a template to amplify the levels of IFITM1, IFITM2, IFITM3, cathepsin B, and NPC1. The final time point (120 h) represents the mRNA levels of that particular gene in what are considered to be differentiated, monocyte-derived DCs. The mRNA levels of each gene were normalized to GAPDH. Each point represents a separate sample evaluated in triplicate. See also Table 2.

Fluorescence microscopy.

Monocytes were infected with green fluorescent protein (GFP)-tagged VLPs pseudotyped with EBGP as described above except that the cells were spinoculated at 1,850 rpm for 45 min at 4°C onto coverslips. Monocyte VLPs were gently washed twice with PBS supplemented with calcium and magnesium (PBS-CM) before incubation at 37°C in 2% RPMI medium for 0, 1.5, and 4 h. The cells were fixed in 4% paraformaldehyde in PBS-CM and cell surface stained with fluorescently labeled wheat germ agglutinin (Invitrogen). Coverslips were washed in PBS-CM twice and once in water and then mounted in antifade reagent (ProLong Gold; Invitrogen) before they were viewed using a confocal Leica SP5 DMI microscope. At least five different random fields were used to determine the number of fluorescent VLPs that were taken up by monocytes.

Measurement of cytokines.

Cytokine measurements were conducted in triplicate using a human cytokine 30-plex panel (Invitrogen) in accordance with the manufacturer's instructions. The data were acquired using a LuminexFlexMAP 3D system (Bio-Rad) and exported to Bio-Rad Bioplex Manager 6.0 for data analysis. Standard curves were optimized by using a software algorithm based on a five-parameter logarithmic curve fit.

RESULTS

EBOV GP-mediated entry into monocytes is restricted.

Several published studies report the productive replication of EBOV in primary human monocytes (4, 20), while others report that EBOV GP pseudotyped viruses are unable to enter monocytes (21–23). To begin to address this apparent discrepancy, we used an established virus-like particle (VLP) entry assay in which VLPs pseudotyped with either EBOV GP or control VSV glycoprotein (VSV-G), either of which can mediate entry (50), contain β-lactamase fused to VP40 (BLA-VP40). Purified VLPs are incubated with a target cell for a defined period of time after which entry is assayed. Successful entry results in the delivery of β-lactamase into the target cell cytoplasm where it cleaves a preloaded cytosolic fluorescent substrate (Fig. 1A). Once cleaved by β-lactamase the substrate-loaded cells fluoresce blue, while uncleaved substrate-loaded cells fluoresce green (50, 58).

Fig 1.

EBOV entry into monocytes is restricted. (A) Schematic representation of the entry assay protocol. β-Lactamase-VP40 (BLA-VP40) containing VLPs pseudotyped with either EBOV GP or VSV G is incubated with APCs for 3.5 h. Penetration into the cytoplasm is detected using a cytosolic substrate that fluoresces blue once cleaved by BLA. Mock and purified EBOV GP VLPs and VSV G VLPs were used in entry assays using human (B) and mouse (C) monocytes, macrophages, and human monocytes previously cultured for 5 days in the presence of M-CSF, M-CSF+ IL-4, M-CSF + TGF-β (TGF-b), M-CSF + IFN-γ (IFN-g) (D) and GM-CSF, GM-CSF+ IL-4, GM-CSF + TGF-β (TGF-b), GM-CSF + IFN-γ (IFN-g) (E).

We first tested whether EBOV VLP entry into freshly isolated undifferentiated monocytes was restricted compared to macrophages and DCs that were differentiated from M-CSF or GM-CSF+IL-4 cytokine-treated monocytes, respectively (54, 59–61). VLPs were normalized for total β-lactamase activity from purified VLP preparations and were used to infect both human (Fig. 1B, D, and E) and mouse (Fig. 1C) monocytes and macrophages. EBOV GP-mediated entry was restricted in human (Fig. 1B) and mouse (Fig. 1C) monocytes. In contrast, equivalent levels of VLPs entered human (Fig. 1B and D) and mouse peritoneal (Fig. 1B) macrophages. Although EBOV GP-mediated entry into monocytes was restricted (consistently ≤10% [data not shown]), control VSV G-mediated entry into human monocytes was efficient (less so in mouse monocytes), demonstrating that the BLA-VP40 assay can detect entry into human (Fig. 1B) and mouse (Fig. 1C) monocytes. Assays testing entry restriction into monocytes compared to DCs were repeated three times and gave results similar to those shown in Fig. 1B and C. Since macrophages can be differentiated from monocytes, these data suggest that entry permissiveness is associated with a differentiated phenotype.

Efficient EBOV entry into differentiated monocytes.

We also found that entry into macrophages differentiated from monocytes by M-CSF treatment was not significantly altered or restricted if we also included IL-4, transforming growth factor β (TGF-β), or gamma interferon (IFN-γ)—cytokines that regulate the expression of proteins, such as DC-SIGN, implicated in EBOV entry (Fig. 1D and data not shown) (62–64). Similarly, entry into human blood monocyte-derived DCs differentiated from monocytes by GM-CSF treatment (Fig. 1E) or mouse bone marrow-derived DCs (not shown) was not significantly affected by IL-4, TGF-β, or IFN-γ treatment.

Kinetics of EBOV entry into differentiating monocytes.

Next, we determined at what point differentiating monocytes become permissive for entry. Monocytes are typically differentiated into immature DCs by incubation with GM-CSF plus IL-4 (54) for 5 to 7 days (50, 51, 55, 56). Therefore, we treated monocytes with these cytokines for periods of 0, 18, 48, 72, 96, and 120 h before testing for entry (Fig. 2A). As seen above (Fig. 1B), VSV G-mediated entry into monocytes was highly efficient and did not require differentiation-inducing cytokine treatment (Fig. 2B). On the other hand, increased EBOV GP-mediated entry efficiency was associated with increased duration of cytokine exposure. At 18 h after cytokine treatment, <10% of the monocytes permitted entry, and only by 48 h was substantial entry into differentiating monocytes detected (Fig. 2B). This experiment was repeated twice with similar results.

Fig 2.

EBOV entry and infection occurs during monocyte differentiation. A schematic representation of the experimental protocol used to test entry after induction of monocyte differentiation is shown in panel A. (A and B) Freshly isolated blood monocytes were cultured in the presence of GM-CSF+IL-4 for 0, 18, 48, 72, 96, and 120 h, harvested and tested for entry using mock, EBOV GP, EBOV GP F88A (entry mutant), or VSV G VLPs. (C and D) EBOV_GFP was used to infect monocytes and DCs at an MOI of 5 and 10. EBOV titers were determined on VeroE6 cells. Cells were harvested at day 2 (C) and day 3 (D) postinfection and subjected to flow cytometry to determine the percentages of cells expressing GFP (EBOV infection).

EBOV infection of monocytes is delayed compared to DCs.

A recombinant EBOV expressing GFP (EBOV_GFP [50]) was used to test infection efficiency in monocytes and DCs. Using EBOV_GFP at MOIs of 5 and 10, monocytes and DCs were incubated with virus for 2 h, washed, and incubated at 37°C. Few cells exhibited GFP fluorescence by 24 h postinfection (data not shown). At 48 h postinfection, there was an ∼3-fold difference in the number of GFP⁺ monocytes and DCs (Fig. 2C), but this difference was no longer apparent by 72 h postinfection (Fig. 2D) when the percentage of infected cells was comparable. The difference in the percentages of GFP⁺ cells indicates that undifferentiated monocytes are permissive for EBOV infection, but the kinetics of replication is slower in monocytes compared to DCs. The 72-h time point from this experiment was repeated, confirming the reproducibility of these results.

Cultured monocytes become permissive for EBOV entry in the absence of supplemental cytokines.

Freshly plated and cultured ex vivo monocytes slowly and spontaneously differentiate (65–67). Therefore, we hypothesized that monocytes become permissive for EBOV entry as they differentiate. We tested the kinetics of EBOV entry when EBOV VLPs were added to plated monocytes and cultured for up to 48 h in media not supplemented with cytokines. In the experiments described above (Fig. 1A), VLPs were incubated with target cells for 3.5 h before testing for cytoplasmic penetration. In this experiment (Fig. 3A), VLPs were added to monocytes or DCs, gently washed, and the VLP target cell mixture cultured for up to 2 days. At 2, 4, 6, 12, 24, and 48 h after the addition of VLPs, monocytes and DCs were tested for entry (Fig. 3B). We observed significant entry into DCs by 2 h after VLP incubation, but significant entry into monocytes did not occur until the 12-h time point. In both cases, the percentage of entry-positive cells reached a plateau at ∼24 h after VLP treatment; however, monocytes never achieved the same level of infection as the DCs. This suggests that entry into monocytes was not only delayed but that entry into undifferentiated monocytes is less efficient than entry into fully differentiated DCs. This experiment has been repeated with similar results.

Fig 3.

Delayed EBOV entry into monocytes compared to DCs. (A) Schematic representation of a modified entry protocol used. Instead of incubating VLPs with target host cells for 3.5 h (as in Fig. 1A), target cells and VLPs were incubated together for 2, 4, 6, 12, 24, and 48 h. The cells were then harvested and tested for VLP entry. (B) Mock (ctl) and purified EBOV GP VLPs were incubated with freshly isolated monocytes and DCs for 2, 4, 6, 12, 24, and 48 h before testing for penetration into the cytoplasm. Shown is the percentage of entry positive cells. (C) Freshly isolated monocytes, 24-h-cultured monocytes and differentiated DCs were stained with antibodies against CD1c, CD14, ICAM-1, and CD44 and assayed by flow cytometry for cell surface expression (red line) compared to background staining (blue line).

Since significant entry into monocytes occurred after 24 h of culture, we tested whether there was a significant difference in the levels of the cell surface differentiation and adhesion markers CD1c, CD14, CD11b, CD11c, ICAM-1, integrin β2, CD44, MHC1, or MHC2 expressed between freshly isolated monocytes and monocytes cultured for 24 h (Fig. 3C). Control DC cells, but not monocytes, expressed significant cell surface levels of DC marker CD1c, whereas the monocytes expressed higher levels of CD14. However, there were no significant differences in the expression levels on freshly isolated monocytes compared to monocytes cultured for 24 h of any of the markers tested, including ICAM-1 and CD44 (data not shown and Fig. 3C).

Because primary monocytes spontaneously differentiate in culture, it is difficult to determine whether differentiation is required for the delayed EBOV VLP entry. As an alternative, the human monocyte-like cell line THP-1 was used in experiments to address the role of differentiation in EBOV infection of monocytes. This cell line is commonly used as an experimental model for human monocytes, but these cells do not spontaneously differentiate in culture. However, THP-1 cells can be induced to differentiate by the addition of phorbol myristate acetate (PMA) (68). This results in an adherent macrophage-like phenotype and renders the cells permissive for EBOV GP-mediated entry (22). We therefore tested for entry into THP-1 cells treated or not with PMA (Fig. 4A). As with primary human monocytes, the undifferentiated THP-1 cells were permissive for entry by VSV G, but not by EBOV GP VLPs, whereas the differentiated THP1s allowed entry (Fig. 4B). We again tested for the expression of CD1c, CD14, CD11b, CD11c, ICAM-1, integrin β2, CD44, MHC1, or MHC2 on untreated and PMA-treated THP-1. Of all the markers tested, only expression of ICAM-1, previously shown to be upregulated on monocytes cultured for several days (69), was significantly upregulated on PMA-treated THP-1 cells (Fig. 4C). The combination of significant ICAM-1 expression and adherence (data not shown) suggests that the THP-1 cells have differentiated into a macrophage-like cell. Since incubating blood monocytes with VLPs for 24 h was sufficient time to allow significant EBOV GP-mediated VLP entry (Fig. 3B), we tested whether a 24-h incubation with VLPs, in the absence of differentiation stimuli, would rescue EBOV GP-mediated entry into THP-1 cells and included primary monocytes as a positive control (Fig. 4D and E). As previously seen (Fig. 3B), VSV G VLPs entered THP-1 cells, whereas the increased VLP incubation time of 24 h did not rescue EBOV GP-mediated entry (Fig. 4E). However, the concomitant addition of PMA and EBOV GP VLPs allowed EBOV GP-mediated entry. VLPs pseudotyped with mutant GP-F159A were previously demonstrated to be entry defective (53, 70). Therefore, GP-F159A VLPs were used in entry assays to exclude the possibility that differentiated monocytes (Fig. 4D) or THP-1 cells (Fig. 4E) might indiscriminately take up any GP-expressing VLPs. The GP-F159A VLPs showed no entry, demonstrating that neither the extended time of VLP-cell incubation nor PMA treatment could render cells permissive for this mutant. Taken together, the data demonstrate that an increased VLP-cell incubation time in blood monocytes but not in THP-1 cells eventually allows EBOV GP entry. THP-1 cells, which do not undergo spontaneous differentiation, allow EBOV GP mediated entry only in the presence of PMA. Therefore, entry into monocytes correlates with their differentiation.

Fig 4.

THP-1 cells become permissive for EBOV entry after PMA-mediated differentiation. (A) Schematic representation depicts the experimental protocol used for experiment shown in panel B. (B) Purified Mock, EBOV GP, or VSV G VLPs were used to test entry permissiveness of THP-1 cells previously treated for 24 h with mock (no Tx) or PMA. Bar graphs show the percentages of entry positive THP-1 cells. (C) THP-1 cells treated or not with PMA for 24 h were stained with antibodies against CD14, ICAM-1, and CD44 and assayed by flow cytometry for cell surface expression (red line) compared to background staining (blue line). (D and E) Mock (no VLPs), EBOV GP F159A (an entry mutant), EBOV GP VLPs, and VSV G VLPs were incubated for 24 h with monocytes (D) and THP-1 cells (E) in culture media supplemented with or without PMA before testing for penetration into the cytoplasm. Shown are the percentages of entry-positive cells.

EBOV transcription is delayed in monocytes compared to DCs.

Previous studies have shown that EBOV productively infects monocytes. We sought to assess the entry kinetics of replication-competent EBOV into undifferentiated monocytes, DCs, undifferentiated THP-1 cells, and differentiated PMA-treated THP-1 cells. As a surrogate measure for viral entry, we profiled EBOV-infected cells for the onset of viral transcription, using quantitative reverse transcription-PCR (RT-PCR) (Fig. 5). Viral transcription was assessed by measuring viral nucleoprotein (NP) mRNA levels at 0, 6, 12, 24, and 48 h after EBOV infection at an MOI of 1. Consistent with the entry assay data, the appearance of EBOV NP mRNA in monocytes was significantly delayed in comparison to infected DCs (Fig. 5A and B), demonstrating that EBOV infection, like EBOV VLP entry, is delayed in undifferentiated monocytes compared to differentiated cells. Furthermore, NP expression was apparent only in differentiated THP-1 cells and not in untreated THP-1 cells, further supporting the view that differentiation is required for early events in EBOV infection (Fig. 5C and D).

Fig 5.

Kinetics of EBOV NP mRNA expression and EBOV GP VLP_GFP uptake in monocytes. EBOV was used to infect—at an MOI of 1—DCs (A), monocytes (B), and THP-1 cells (C) or THP-1 cells grown in medium supplemented with PMA (D). RNA was harvested from cells at 0, 3, 6, 12, 24, and 48 postinfection. Poly(A)-derived cDNA was used as a template for quantitative RT-PCR analysis of EBOV nucleoprotein (NP) mRNA. The relative copy number is shown. (E) EBOV GP VLP_GFPs (green) were incubated with monocytes for 0, 1.5, and 4 h before fixation. Cell plasma membranes were stained (red) using wheat germ agglutinin conjugated to a fluorochrome. Uptake of VLPs was analyzed using confocal microscope Leica SP5 DMI. (F) The percentages of EBOV GP VLP_GFP were scored as entering beyond the cell periphery (taken up) by monocytes if the VLP was inside the cell and not associated with the membrane. Totals of 39, 351, and 21 VLPs were scored at time points 0, 1.5, and 4 h, respectively.

EBOV GP VLP_GFP uptake in monocytes and DCs.

To assess the kinetics of EBOV GP VLP uptake into monocytes, we used fluorescently tagged VLPs that were generated by coexpression of a chimeric GFP-VP40 protein and GP, leading to the formation of EBOV GP VLP_GFP (51). Monocytes were treated with equivalent amounts of EBOV GP VLP_GFPs, washed gently, and incubated for 0, 1.5, and 4 h to allow the uptake of particles. Using confocal microscopy, VLP uptake was assessed by visualizing the location of EBOV GP VLP_GFPs relative to the monocyte surface membrane. At 0 and 1.5 h postinfection, there was a clear association of fluorescently labeled cell membranes (red) with fluorescent particles (green) (Fig. 5E). The VLPs remained associated with the cell periphery until 4 h, when significant numbers of the EBOV GP VLP_GFPs had relocalized toward the interior of the monocytes (Fig. 4E). Specifically, at 0, 1.5, and 4 h postinfection, 0, 5, and 48% of EBOV GP VLP_GFPs, respectively, were in the cell interior regions of monocytes (Fig. 5F). These data, which have been reproduced, demonstrate that EBOV GP VLP_GFPs readily associate with monocytes before entry occurs.

Monocytes secrete limited levels of cytokines upon EBOV infection.

Cytokines produced by APCs are proposed to contribute to EBOV pathogenesis (18). To test for the presence and levels of several inflammatory cytokines, a multiplex assay was performed on supernatants harvested 3 days after EBOV infection of monocytes, macrophages, and DCs. A representative panel of secreted cytokines (IL-12, IL-6, IL-8, MIP-1α, MCP-1, RANTES, IFN-α, IP-10 and G-CSF) is presented in Fig. 6. Consistent with previous studies, macrophages in general secreted comparably more inflammatory cytokines, but monocytes did secrete some inflammatory cytokines, albeit at lower levels. Therefore, the differentiation status and the type of APC affects both the efficiency/kinetics by which EBOV infection is established and the inflammatory cytokine response.

Fig 6.

Secreted cytokines induced by EBOV infection of PBMCs, monocytes, macrophages, DCs. PBMCs, monocytes, monocyte-derived macrophages, or DCs were mock infected or infected with EBOV at MOI of 5 (the cell type is identified with a label under the first of the mock and infected histograms). At 3 days postinfection, the concentrations (pg/ml) of IL-12, IL-6, IL-8, MIP-1α, MCP-1, RANTES, IFN-α, IP-10, and G-CSF in spent supernatants were determined by using a human 30-plex cytokine assay.

Expression levels of host factors critical for entry change during monocyte differentiation and EBOV infection.

Since significant changes occur in the monocyte transcriptome during differentiation (71), we tested the hypothesis that the levels of host factors implicated in EBOV entry also change during monocyte differentiation. The relative mRNA levels of select host entry factors (Table 2) were compared by quantitative RT-PCR between undifferentiated monocytes and DCs. Of the mRNAs examined, only the mRNA levels of cathepsin B, NPC1, IFITM1, and IFITM3 showed a >10-fold difference between monocytes and DCs. Therefore, these specific mRNAs were chosen for further investigation. The kinetics of mRNA expression of these host factors was determined in cultured monocytes at 0, 3, 6, 12, 24, and 48 h after plating. For comparison, the 120-h time point represents the relative mRNA levels in DCs. We also tested for the expression of cathepsin B, NPC1, IFITM1, IFITM2, and IFITM3 mRNA levels after monocyte infection with EBOV at an MOI of 1 (Fig. 7). Comparison of undifferentiated monocytes with or without EBOV infection revealed several striking findings. Although IFITM entry restriction factor expression in cultured monocytes dramatically decreases over time to the levels measured in DCs, their expression initially increases during the first few hours after plating. However, upon EBOV infection of monocytes, IFITM expression immediately decreases. On the other hand, these same infected cells demonstrate an upregulation of factors required for entry, namely, NPC1 and cathepsin B. These data demonstrate that the levels of host entry factors not only change during monocyte differentiation, but EBOV infection also hastens the downregulation of entry restriction factors, while at the same time upregulating host factors that are required for entry.

Table 2.

RelativemRNA levels of EBOV entry factors expressed by monocytes and DCs

Entry factor	Relative monocyte/hDC expressiona
Located at the cell surface
TIM-4	ND in monocytes
MER	Similar
DTK	Monocytes > DCs
AXL	Monocytes > DCs
ITGB1	Similar
ITGB2	Similar
hMGL	Similar
LSECTin	ND
L-SIGN	ND
DC-SIGN	Similar
Located intracellularly
NPC1	Monocytes < DCs*
VPS33A	Similar
VPS11	Similar
VPS18	Similar
VPS41	Monocytes < DCs
BLOC1S1	Similar
GNPTAB	Monocytes > DCs
Cat B	Monocytes < DCs*
IFITM1	Monocytes > DCs*
IFITM2	Monocytes > DCs
IFITM3	Monocytes > DCs*

^aThe relative levels of EBOV entry factors were assayed for in freshly isolated monocytes, as well as in DCs, and normalized to the levels of GAPDH. It is indicated, for each gene, was whether the gene was expressed at a higher (>) or lower (<) level in monocytes than in DCs. ND, not detected.

^*, A >10-fold difference in mRNA levels.

Stable expression of NPC1 from THP-1 cells partially rescues EBOV entry.

NPC1 is essential for EBOV infection, and there was a 100-fold increase in NPC1 mRNA in DCs and macrophages relative to monocytes (Fig. 8A). Western blotting for the levels of NPC1 confirmed that the lower levels of mRNA correspond to lower levels NPC1 protein expression in monocytes and higher levels in DCs and macrophages (Fig. 8B). PMA differentiated THP-1 cells also demonstrated higher levels of NPC1 protein compared to their undifferentiated counterparts (Fig. 8B). To test whether overexpressing NPC1 can rescue EBOV entry into undifferentiated monocytic cells, THP-1 cells were stably transduced with a retrovirus that expresses NPC1 and tested for EBOV GP VLP entry. Overexpressing NPC1 partly rescued EBOV GP-mediated entry (Fig. 8C), suggesting that low levels of NPC1 in monocytes may determine, in part, their resistance to EBOV entry. This experiment was repeated with similar results.

Fig 8.

Stable expression of NPC1 in THP-1 cells rescues EBOV GP VLP entry. (A) Relative NPC1 mRNA levels in monocytes, macrophages, and DCs as normalized to GAPDH mRNA. (B) Western blots of the levels of NPC1 expressed in monocytes, DCs, macrophages, THP-1 monocytes, and THP-1 monocytes pretreated with PMA. Also shown are the protein levels of housekeeping gene GAPDH. (C) Relative entry of mock and EBOV GP VLPs into THP-1 cells, THP-1 cells pretreated with PMA, and THP-1 cells stably expressing NPC1.

DISCUSSION

This study addresses seemingly contradictory observations that monocytes are productively infected by EBOV but are nonpermissive for EBOV GP-mediated entry (4, 20–23). Our initial studies, wherein VLPs were incubated with undifferentiated or differentiated target cells for 3.5 h (Fig. 1A), indicated that in the presence of GM-CSF+IL-4 (a standard protocol used to generate monocyte-derived DCs), monocytes become permissive for entry by 48 h after cytokine addition (Fig. 2A and B). On the other hand, if VLPs were incubated for longer periods of time with cultured monocytes, which spontaneously differentiate, substantial entry was detected as early as 24 h post-VLP addition (Fig. 3B), although overall entry efficiency was consistently lower compared to DCs (Fig. 3B). Studies with infectious EBOV, where we used NP mRNA expression as a surrogate marker for entry (Fig. 5), support this conclusion. These data suggest that monocytes are not incompetent for EBOV entry. Rather, entry is substantially delayed and somewhat less efficient compared to DCs.

Primary monocytes spontaneously differentiate in culture. Therefore, it is problematic to determine whether or not the delay in entry requires the differentiation process. We therefore turned to THP-1 cells, because these cells do not spontaneously differentiate in culture, exhibit a strict resistance to entry as monocytes, and become permissive following differentiation overnight with PMA (Fig. 4B) (22). In these cells, prolonged incubation with EBOV VLPs, in the absence of PMA did not lead to successful entry. However, if prolonged incubation occurred with the coaddition of PMA, then THP-1 cells became entry positive by 24 h postinfection (Fig. 4E). Although we cannot fully exclude the possibility that entry requirements differ in THP-1 cells compared to primary monocytes, these results suggested that differentiation is a requirement for EBOV entry into monocytes.

Significant changes occur in the monocyte transcriptome during differentiation (71), and this leads to altered expression of select cell surface markers. Although no significant differences were seen in the expression levels of such cell surface markers when comparing freshly isolated monocytes and 24-h-cultured monocytes, which are entry permissive (Fig. 3C), monocyte differentiation and entry permissiveness did correlate with a decrease in mRNA levels for entry restriction factors and an upregulation of factors critical for EBOV entry, all of which occurred within hours of monocyte culturing. Therefore, it is likely that monocytes become permissive for entry before the differentiation process is complete. Among the entry-relevant factors that change, IFITM1, IFTIM2, and IFITM3 are among a family of IFN-inducible proteins that when overexpressed restrict the entry of a variety of viruses, including EBOV (47, 48, 72–75). We observed a dramatic downregulation of the mRNA for each of these proteins during the differentiation process, and this downregulation was accelerated in EBOV-infected monocytes. Conversely, essential entry factors cathepsin B and NPC1 are dramatically upregulated during monocyte differentiation. In THP-1s, the patterns of NPC1 expression mirror what was seen in primary monocyte to macrophage or DC differentiation, and the expression of NPC1 via a retroviral vector was sufficient to partially rescue entry into undifferentiated THP-1 cells (Fig. 8). Whether NPC1 facilitates entry directly, by acting as a receptor, or by some indirect mechanism remains to be determined. Nonetheless, these data point to NPC1 upregulation as a critical determinant of entry into APCs. Whether coexpression of other entry factors (e.g., cathepsin B) or simultaneous knockdown of restriction factors (IFITMs) would further enhance entry remains to be determined. In addition to their role in EBOV entry, it will be of interest to determine how the regulation of these factors generally affects monocyte and macrophage function.

Our microscopy studies indicate that VLPs attach to undifferentiated monocytes at some frequency and that the VLPs can remain associated with the cells for at least several hours. However, by the 4-h time point when the β-lactamase-based entry assay does not detect significant entry (Fig. 3) and NP mRNA remains undetectable in EBOV-infected monocytes (Fig. 5), a significant percentage of the particles had moved to an intracellular location (Fig. 5). It remains to be determined whether the inward migration of VLPs at 4 h postinfection represents a pathway that leads to productive entry. However, the data suggest that monocytes are capable of supporting a certain level of viral attachment and that VLPs remain monocyte-associated as these cells undergo differentiation and become fully permissive for entry.

Previous studies implicated cell adherence versus nonadherence as a determinant of EBOV entry into THP-1 and 293 cells (22). Because differentiation of primary monocytes is intertwined with the acquisition of adherence, we cannot exclude the possibility that the modulation of entry factors is a direct consequence of adherence. Comparison of the ability of nonadherent and adherent cells, including monocytes and macrophages, to bind to a recombinant EBOV GP receptor binding domain (RBD)-Fc protein also indicated that adherence correlates with the translocation of an RBD binding activity from an intracellular compartment to the cell surface (21). Although our studies did demonstrate stable association of VLPs with freshly isolated, nonadherent primary monocytes, these studies do not directly address the presence of an RBD binding factor on monocytes versus macrophages.

There are multiple potential consequences for the delayed entry of EBOV into monocytes. Monocytes are blood-borne phagocytic mononuclear cells (76) that can capture and present antigen. Under inflammatory conditions, monocytes can also enter tissues in order to differentiate into macrophages or DCs (77, 78). The persistence of EBOV GP VLPs associated with monocytes suggests the possibility that EBOV can hijack monocytes and potentially “hide” from the host immune response within monocytes, using migrating monocytes as a vehicle for dissemination. In this model, completion of the entry process and onset of viral replication might only occur after the monocytes enter tissue and differentiate. In addition, we show that EBOV-infected monocyte cultures secrete limited levels of cytokines (Fig. 6). Further studies are required to determine whether the lack of robust secretion of a subset of cytokines is related to the delayed kinetics of viral entry.

EBOV infection appears to accelerate the transcriptional regulation of entry-relevant host factors compared to uninfected, plated monocytes (Fig. 7). EBOV can trigger signaling in macrophages within 1 h postinfection (79) and EBOV VLPs, which lack viral replication machinery can also activate macrophages as well as DCs. Furthermore, several of these studies have shown that EBOV VLPs stimulate APCs in a GP-dependent manner, illustrating a central role for the attachment protein in an entry process that itself induces signaling (3, 51, 80). Therefore, it will be of interest to identify the signals responsible for EBOV-induced transcriptional regulation. By defining the pathways required for EBOV to differentiate and enter monocytes, it may be possible to devise strategies that will impair the infection of APCs, potentially blunt their cytokine and other responses and thereby influence EBOV pathogenesis.